Multiple genetic modifications of the erythromycin polyketide synthase to produce a library of novel ‘ ‘ unnatural ’ ’ natural products

The structures of complex polyketide natural products, such as erythromycin, are programmed by multifunctional polyketide synthases (PKSs) that contain modular arrangements of functional domains. The colinearity between the activities of modular PKS domains and structure of the polyketide product portends the generation of novel organic compounds—‘‘unnatural’’ natural products—by genetic manipulation. We have engineered the erythromycin polyketide synthase genes to effect combinatorial alterations of catalytic activities in the biosynthetic pathway, generating a library of >50 macrolides that would be impractical to produce by chemical methods. The library includes examples of analogs with one, two, and three altered carbon centers of the polyketide products. The manipulation of multiple biosynthetic steps in a PKS is an important milestone toward the goal of producing large libraries of unnatural natural products for biological and pharmaceutical applications. Few molecules have captured interest in both chemotherapy and chemistry to the extent of the polyketide erythromycin. Erythromycin and its semisynthetic derivatives are the third most widely used class of antibiotics, with current worldwide sales exceeding 3.5 billion dollars. In addition, erythromycin analogs are gaining interest for their potential use in the treatment of gastrointestinal disorders and inf lammatory diseases and as next-generation antibiotics for treatment of emerging drug-resistant strains of bacteria (1, 2). The chemical challenges of erythromycin attracted the talents of R. B. Woodward and 48 colleagues who completed its synthesis in 1981 (3) and of a cadre of medicinal chemists who prepared analogs leading to the important second generation of macrolide antibiotics—clarithromycin, azithromycin, and others (4). Although such efforts effectively saturated the chemical modifications possible at the existing functional groups of the macrolide ring, most of the ring remained inaccessible to chemical modification. After the discovery of the programmed nature of modular polyketide biosynthesis (5, 6), genetic engineering strategies emerged for the production of novel polyketides (7, 8). With the work described here, it is now feasible to contemplate modifications of the chemically intractable sites of such molecules to produce hundreds or even thousands of new ‘‘unnatural’’ natural products by genetic engineering. Such novel macrolides could in themselves provide the basis for new pharmaceuticals or could serve as scaffolds for new semisynthetic analogs. It is now well established that the ‘‘modular’’ polyketide synthases (PKSs) (i) each are encoded by a cluster of contiguous genes and (ii) have a linear, modular organization of similar catalytic domains that both build and modify the polyketide backbone. Each module contains a set of three domains—a ketosynthase, an acyltransferase (AT), and an acyl carrier protein (ACP)—that catalyze a 2-carbon extension of the growing polyketide chain (Fig. 1) (9). The choice of extender unit used by each module—acetate, propionate, or other small organic acids in the form of CoA thioesters—is determined by the specificity of the AT domain (10–12). With each 2-carbon chain extension, the final state of the b-carbonyl is embedded as a ketone, hydroxyl, methenyl, or methylene group by the presence or absence of one, two, or three additional catalytic domains in the module—a ketoreductase (KR), dehydratase (DH), and enoyl reductase (ER). In effect, the composition of catalytic domains within a module provides a ‘‘code’’ for the structure of each 2-carbon unit, and the order of modules codes for the sequence of the 2-carbon units, together creating a template colinear with the polyketide product. The remarkable structural diversity of polyketides is governed by the combinatorial possibilities of catalytic domains within each module, the sequence and number of modules, and the post-polyketide synthesis cyclization and ‘‘tailoring’’ enzymes that accompany the PKS genes. The direct correspondence between the catalytic domains of modules in a PKS and the structure of the resulting biosynthetic product portends the possibility of modifying polyketide structure by modifying the domains of the modular PKS. We have constructed a combinatorial library of polyketides by using 6-deoxyerythronolide B synthase (DEBS), the PKS that produces the macrolide ring of erythromycin. This was accomplished by substituting the ATs and b-carbon processing domains of DEBS with counterparts from the rapamycin PKS (RAPS) (13) that encode alternative substrate specificities and b-carbon reductionydehydration activities. Engineered DEBS containing single, double, and triple catalytic domain substitutions catalyzed production of erythromycin macrolactones with corresponding single, double, and triple modifications. The ability to simultaneously manipulate multiple catalytic centers of the PKS demonstrates the robustness of the engineering process and the potential for creating libraries of novel polyketides that are impractical to prepare in the chemistry laboratory. MATERIALS AND METHODS Strains, Culture Conditions, and DNA Manipulation. DNA manipulations were performed in Escherichia coli XL1 Blue The publication costs of this article were defrayed in part by page charge payment. This article must therefore be hereby marked ‘‘advertisement’’ in accordance with 18 U.S.C. §1734 solely to indicate this fact. PNAS is available online at www.pnas.org. Abbreviations: PKS, polyketide synthase; AT, acyltransferase; KR, ketoreductase; DH, dehydratase, ER, enoylreductase; ACP, acyl carrier protein; DEBS, 6-deoxyerythronolide B synthase; RAPS, rapamycin polyketide synthase; 6dEB, 6-deoxyerythronolide B. †Present address: Department of Genetics, Faculty of Science, Kasetsart University, Bangkok 10900, Thailand. *To whom reprint requests should be addressed. e-mail: mcdaniel@ kosan.com.